Multi-agent AI Policy Design Assessment
Trustero Intelligence now comes with built-in playbooks for common GRC tasks and multi-agent capabilities, saving you time and increasing the accuracy of your work.
Trustero Intelligence (TI) introduces conversational AI user interface designed for complex GRC use-cases. TI recognizes your intentions and finds the AI agent that can bring the highest degree of accuracy to the task at hand. Tasks range from conducting simple research of your GRC posture to completing a policy design assessment against a new regulation or framework.
TI is contextualized using your policies, controls, evidence, risk registers, knowledge bases, and more. Trustero can be configured to continuously enrich the GRC data context from a variety of sources including wide ranging SaaS services, GRC platforms, shared drives, workflow ticketing systems, or custom API endpoints. The GRC data context allows TI to create tight contexts for LLMs reducing hallucinations while increasing the citation accuracy.
A Tour of the new Trustero Intelligence (TI)
Meet the new Trustero Intelligence (TI). TI can be launched using the new launch button below the Trustero logo. Once you’re in TI, you’ll have access to the prompt composer where you converse with TI. A tool bar on the upper right of the TI conversational interface provides access to historical chats as well as a GRC context in which TI derives its answers. TI also comes with a set of useful examples called playbooks. These playbooks provide a range of simple to complex tasks TI can perform.
.png)
The TI tool bar provides access to common chat and context management tools. Chat history allows you to refer back to a previous chat. Chats can be public or private to you. Public chats can be shared between users in your Trustero account using the chat’s permalink. Private chats can only be accessed by you. Trustero users use private chats for exploration and to test reusable playbooks. Once reusable playbooks are ready for team consumption, the private chat can be made public.
Each chat session has its own chat context. A chat context consists of GRC data ingested by Trustero. By default, the chat context includes the entire GRC context. Broad GRC context allows one to conduct broad research across the entire context. You can also narrow the context if you have a specific task, such as inspecting a specific policy for errors. Selecting the chat context button in the TI toolbar within a chat session allows you to configure the chat context.
The context includes knowledge base documents uploaded by you or your teammates; compliance related information such as policies, controls, and related evidence; risk related information such as risk registers and 3rd party vendors; and Trustero integrations.
A chat is initiated by selecting one of the playbooks or by using the TI chat composer. Once a chat is initiated, TI switches into the chat session window where you converse with TI. All prompts sent to TI first undergo a planning process. The TI planner will describe how it intends to fulfill your request in a plan, then execute the plan. Certain plans may trigger additional AI agent interaction or AI tool invocations behind the scenes. If an AI agent requires additional details to fulfill its task, you will be prompted to provide the details.
TI Built-in Playbooks
TI provides a dozen playbooks. A playbook is a simple or complex prompt sequence instructing TI to complete a GRC task and generate a report. TI playbooks extend the out-of-the-box capabilities of Trustero. One playbook in particular best illustrates multi-agent TI’s capabilities: The policy design assessment playbook.
Policy Design Assessment
TI provides the ability to assess your policies against any regulation or framework. A multi-agent system working in concert with a planner breaks down the complex task of assessing any or all policies in your account against a regulation such as FFIEC or a framework such as ISO 27001.
Starting with the TI conversational user interface, you begin by either selecting the Policy Design Assessment (PDA) playbook or enter “Do a design assessment of my policies” in the TI prompt composer.
TI’s planning agent will break down the problem into multiple steps, starting with clarifying which policies and frameworks you would like to perform the assessment on.
Once TI knows which policies and framework it needs to perform the assessment, it will send the request to multiple AI agents. The first agent will extract the sections, if there are sections, and extract requirements from each section. A conversion agent will then convert each requirement into multiple questions to ask of your policies.
A questionnaire agent will ask the questions to relevant policies and formulate a conclusion on whether the policies affirmatively answers the question. An analysis agent will determine if the answers adequately fulfill the requirements and cite its sources.
Finally, a reporting agent will generate a gap assessment report detailing the findings. The complete policy design assessment is conducted by AI agents with minimal human input.
Reports generated by AI agents will also be available in a new report section of Trustero. The Reports section provides a list of all agent generated reports.
TI’s multi-agent capabilities can break down a broad request into multiple steps AI agents can complete with a high degree of accuracy saving you time while increasing the accuracy of your work.
Try it Today
The new Trustero Intelligence and Policy Design Assessment agents are available today. Log into your Trustero account to explore the capabilities of multi-agent AI for GRC and policy design assessment.